Researchers at Baylor College of Medicine have presented an artificial intelligence system for diagnosing placenta accreta spectrum (PAS) from standard ultrasound images. PAS is one of the most dangerous complications of pregnancy, in which the placenta abnormally invades the uterine wall, increasing the risk of bleeding during childbirth.
The algorithm was tested on 113 pregnant women at high risk of PAS who were followed at Texas Children’s Hospital between 2018 and 2025. The average gestational age at examination was about 31 weeks. The system analyzed two-dimensional ultrasound scans and identified all confirmed PAS cases.
Two false-positive results were recorded, but no real case of the complication was missed. Such a tool may help doctors plan births in advance and reduce the risks of severe maternal complications.
In recent years, artificial intelligence has been widely adopted in radiology and cardiology, but obstetrics has traditionally been considered a more manual and clinically intuitive field. The emergence of an algorithm that analyzes ordinary two-dimensional ultrasounds — without complex and expensive additional technologies — means it could be scaled up and made relatively accessible. This is especially important for countries where resources are limited and maternal mortality remains high.
